Implement a perceptron in python from scratch and use your implementation to show that the following boolean functions are learnable: i) the NOT function, ii) the AND function, iii) the OR function, and vi) the NAND function. That is, in each case, train a perceptron model on the appropriate data/labels, depict each boolean function in the appropriate data/label space, and depict the decision boundary showing that each of these functions is learnable by a perceptron. Finally, train a perceptron model on the XNOR function and depict this function in the appropriate data/label space along with its decision boundary in order to confirm that the function is not learnable by a perceptron. Now, pretend that no one else has ever seen these results before. Add some comments to your code that explains these depictions.